In thirty years of reviewing fixed-income allocations for single-family offices, I have seen the same mistake repeated with impressive consistency: clients reach for a familiar yield without asking whether the underlying credit has structurally changed. The tax exemption is real, the retail demand is deep, and the inertia is powerful. California municipal bonds are testing all three of those assumptions right now.
The bull case is well known. California is the world’s fourth-largest economy with a GDP exceeding $4.1 trillion. The double federal-and-state tax exemption translates to a combined marginal rate approaching 54.1% for top-bracket California residents, making even a modest nominal coupon genuinely competitive. Morgan Stanley’s municipal team entered 2026 noting that 20-year AA-rated Munis were offering taxable-equivalent yields of just under 7%. On paper, the asset class earns its allocation.
But BlackRock’s municipal credit team offered a more candid read in August 2025: strong retail demand has produced spreads that “traditionally do not reflect the fundamental picture.” That is institutional language for investors not being paid for the risk they carry. I agree.
The Pension Overhang No One Wants to Price
The Reason Foundation’s December 2025 report places California’s combined unfunded pension liabilities — across CalPERS, CalSTRS, and local systems — above $265 billion under standard actuarial assumptions. CalPERS alone accounts for $166 billion. Apply the market-based discount rate that Stanford’s Joshua Rauh at the Hoover Institution advocates, risk-free rates instead of the assumed-return fiction of 6.8%, and the gap widens materially beyond what state accounting shows.
The operational consequences are visible. Pension contributions grew from 2.1% of the state budget in 2002–2003 to over 6.5% by 2016–2017, and the trajectory has not reversed. Los Angeles and San Jose now direct more than 10% of their budgets to pension costs alone. CalPERS has doubled its private equity allocation, from 6% to 17% of assets, and is targeting 40% in private assets to chase the returns it cannot reliably achieve in public markets. That is a high-fee, illiquid bet made with taxpayer-guaranteed obligations as the backstop. If those return targets are missed, contribution demands on already-strained general funds increase.
Migration Is a Balance Sheet Problem
California’s debt service is structurally dependent on personal income and capital gains tax revenue generated by a narrow band of mobile, high-earning residents. U-Haul’s 2024 Growth Index showed California leading the nation in outbound one-way moves for six consecutive years, with high-profile corporate relocations to Texas, Arizona, and Nevada following the same vector. When the top 1% of filers generate roughly 40% of personal income tax receipts, the exit of a cohort of technology executives is not a demographic footnote, it is a credit event in slow motion.
Governance Risk Is Underpriced
My experience as an expert witness in fiduciary disputes has sharpened my reading of California’s recent control failures. The state’s Employment Development Department disbursed an estimated $20 billion in fraudulent or improper unemployment claims during the pandemic, recovering roughly $6 billion to date. The SBA suspended over 111,000 California borrowers linked to $8.6 billion in PPP and EIDL fraud. In April 2026, prosecutors dismantled a Medi-Cal hospice scheme involving $267 million in phantom services and shell entities. The high-speed rail project now carries a $126 billion price tag, multiples more than its original voter-approved estimate. These episodes are not anomalies, they are symptoms of control environments that would fail a basic institutional audit. For a bondholder, each dollar diverted through fraud or mismanagement is a dollar unavailable for debt service. At scale, that is a structural credit risk, not an operational nuisance.
What the Pricing Is Telling You — and What It Isn’t
None of this is hidden. The Equable Institute publishes state-by-state pension data. The Reason Foundation’s research is publicly available. Yet California spreads remain compressed because technical factors — a captive retail base, the tax exemption, and Muni fund inflows of $50 billion in 2025 — consistently overwhelm fundamental credit analysis in the short run. This is not the first time that technical demand has dominated credit pricing longer than logic would predict. It is, however, a dynamic that historically resolves through spread widening when a specific catalyst forces the market to reprice. Detroit carried a Moody’s AA rating eighteen months before it filed for Chapter 9.
I am not predicting California default. The relevant question for a family office trustee is narrower: does the current yield adequately compensate for $265 billion in off-balance-sheet pension obligations, a shrinking high-income tax base, documented governance failures at scale, and a political structure with limited near-term incentive to reform? My answer, formed over three decades of advising UHNW families, is no.
Portfolio Implications
These are not abstract concerns. They have direct implications for how a trustee should be positioned today.
Three adjustments merit consideration. First, reduce long-duration California GO exposure in favor of shorter-duration essential-service revenue bonds, BlackRock’s municipal team draws the same distinction, preferring revenue bonds over GOs within the state. Second, apply single-state concentration limits: a 5% allocation in a nationally diversified portfolio is defensible; 20% in a California-concentrated family is not. Third, stress-test holdings against 150–200 basis points of spread widening — a conservative assumption by historical credit-cycle standards — to quantify mark-to-market exposure before it becomes a realized loss.
The tax efficiency of California Munis is real. So are the structural risks. What the current spread does not provide is adequate compensation for both. Fiduciary duty requires asking that question even when the market has not yet bothered to answer it.
